High-speed microscopy finds three of four sugar-building enzymes have moving domains

The team did not lean on the microscope alone. Mass photometry assessed the assembly states of all four proteins ^[5], solution measurements helped establish how they assembled ^[7], and a newly determined crystal structure of POMGNT2 gave a reference for interpreting its observed shape ^[8]. That reference matters most for the rigid dimer. A protein that barely moves is informative only if the shape on the surface can be checked against something independent.

Counting by enzyme, three of the four proteins showed flexible domains ^[17]. That is a sample of four. The particles were adsorbed onto a supporting surface in liquid ^[6], and the summary does not report how many were tracked per enzyme or how the surface may influence their motion.

The researchers call K4CP's shift toward compact conformations with chondroitin oligosaccharides a qualitative tendency ^[12]. L137 is the thinnest data point: a glycosyltransferase candidate from the giant virus Mimivirus, analyzed in what the team describes as exploratory work ^[4].

The authors propose that a rigid scaffold could help an enzyme recognize its substrate precisely, and that mobile domains could make room for a growing glycan chain ^[18]. We think that is a reasonable hypothesis. The thing this doesn't tell you is whether rigidity is what gives POMGNT2 precise substrate recognition. The rigid end of the spectrum has a single member ^[9], so its rigidity cannot be separated from whatever else sets that enzyme apart. The four proteins also come from different sources: two human, one bacterial and one viral ^[3][4].